Supplying an Argument in a Function

The following are acceptable arguments for a function:

• Numeric constant, such as 6 or 15.
• Date constant, such as 022802.
• Date in alphanumeric, numeric, date, or AnV format.
• Alphanumeric literal, such as STEVENS or NEW YORK NY. A literal must be enclosed in single quotation marks.
• Number in alphanumeric format.
• Field name, such as FIRST_NAME or HIRE_DATE. A field can be a data source field or temporary field. The field name can be up to 66 characters long or a qualified field name, unique truncation, or alias.
• Expression, such as a numeric, date, or alphanumeric expression. An expression can use arithmetic operators and the concatenation sign (|). For example, the following are valid expressions:

  CURR_SAL * 1.03

  and

  FN || LN

• Dialogue Manager variable, such as &CODE or &DDNAME.
• Format of the output value enclosed in single quotation marks.
• Another function.
• Label or other row or column reference (such as R or E), or name of another RECAP calculation, when the function is called in an FML RECAP command.

Depending on the function, an argument can be in alphanumeric, numeric, or date format. If you supply an argument in the wrong format, you will cause an error or the function will not return correct data. The following are the types of argument formats:

• Alphanumeric argument. An alphanumeric argument is stored internally as one character per byte. An alphanumeric argument can be a literal, an alphanumeric field, a number or date stored in alphanumeric format, an alphanumeric expression, or the format of an alphanumeric field. A literal is enclosed in single quotation marks, except when specified in operating systems that support Dialogue Manager RUN commands (for example, -MVS RUN).
• Numeric argument. A numeric argument is stored internally as a binary or packed number. A numeric argument includes integer (I), floating-point single-precision (F), floating-point double-precision (D), and packed decimal (P) formats. A numeric argument can be a numeric constant, field, or expression, or the format of a numeric field.

  All numeric arguments are converted to floating-point double-precision format when used with a function, but results are returned in the format specified for the output field.

  Note: With CDN ON, numeric arguments must be delimited by a comma followed by a space.

• Date argument. A date argument can be in either alphanumeric, numeric, or date format. The list of arguments for the individual function will specify what type of format the function accepts. A date argument can be a date in alphanumeric, numeric, or date format; a date field or expression; or the format of a date field.

  If you supply an argument with a two-digit year, the function assigns a century based on the YRTHRESH and DEFCENT parameter settings.

An argument is passed to a function by reference, meaning that the memory location of the argument is passed. No indication of the length of the argument is given.

You must supply the argument length for alphanumeric strings. Some functions require a length for the input and output arguments (for example, SUBSTR), and others use one length for both arguments (for example, UPCASE).

Be careful to ensure that all lengths are correct. Providing an incorrect length can cause incorrect results:

• If the specified length is shorter than the actual length, a subset of the string is used. For example, passing the argument 'ABCDEF' and specifying a length of 3 causes the function to process a string of 'ABC'.
• If the specified length is too long, whatever is in memory up to that length is included. For example, passing an argument of 'ABC' and specifying a length of 6 causes the function to process a string beginning with 'ABC' plus the three characters in the next three positions of memory. Depending on memory utilization, the extra three characters could be anything.

Some operating system routines are very sensitive to incorrectly specified lengths and read them into incorrectly formatted memory areas.

The number of arguments required varies according to each function. Functions supplied by Information Builders may require up to six arguments. User-written subroutines may require a maximum of 200 arguments including the output argument. If a function requires more than 200 arguments, you must use two or more calls to pass the arguments to the function.

Arguments must be specified in the order shown in the syntax of each function. The required order varies according to the function.

The USERFCHK setting controls the level of verification applied to DEFINE FUNCTION and Information Builders-supplied function arguments. It does not affect verification of the number of parameters; the correct number must always be supplied.

USERFCHK is not supported from Maintain Data.

Functions typically expect parameters to be a specific type or have a length that depends on the value of another parameter. It is possible in some situations to enforce these rules by truncating the length of a parameter and, therefore, avoid generating an error at run time.

The level of verification and possible conversion to a valid format performed depends on the specific function. The following two situations can usually be converted satisfactorily:

• If a numeric parameter specifies a maximum size for an alphanumeric parameter, but the alphanumeric string supplied is longer than the specified size, the string can be truncated.
• If a parameter supplied as a numeric literal specifies a value larger than the maximum size for a parameter, it can be reduced to the proper value.